Image Quality Decrease and pixelated

The quality of the image decrease when I read and display out in matlab, please advice.
The below are the commands that I had used: imread,imshow
Image Info: Format:JPEG Dimension:1,434 x 1,014 Size:487KB

1 Comment

Try using image or imagesc instead, depending if you have full R, G, B data or just indexed data.
I never use imshow as it always seems to give appalling results.

Sign in to comment.

 Accepted Answer

I don't know what Adam does with imshow() but it works fine for me and is my preferred method, unlike imagesc() which applies some colormap that you didn't ask for. Don't use pcolor - that's even worse. Anyway, why do you say the image quality decreases. Show me a screenshot of the image in imshow() and in some other program (like paint or Photoshop) so I can see why you think it's worse. It should display it exactly as it was stored on disk, just magnified or demagnified for display of course. Of course you may have stored a horrible image where you jpegged the heck out of it and it shows up with horrendous jpeg block artifacts, but that would be present in any program that displays it.

11 Comments

Ian
Ian on 2 Dec 2014
Moved: DGM on 5 Feb 2023
Image 1.jpg is the actual image, Image 2.jpg is from matlab. Thanks for your advices.
It looks like maybe the image is real big and it's being subsampled for display and dropping some lines. You can display full size, pixel for pixel. See attached demo for imscrollpanel.
Ian
Ian on 2 Dec 2014
Moved: DGM on 5 Feb 2023
Thanks for your reply, lets say I want to plot the image together with the graph what can I do to retain the quality?
Image Analyst
Image Analyst on 2 Dec 2014
Moved: DGM on 5 Feb 2023
I don't know exactly. Try export_fig() or change your renderer (painter, zbuffer, openGL) or call the Mathworks and tell them you're not happy with the quality of their graphics.
Ian
Ian on 2 Dec 2014
Moved: DGM on 5 Feb 2023
thanks
Hello,
I know this is a (very) old thread, but I am having an issue currently where it seems some of the picture is shown differently with imshow() than when brought up on another program. _DSC1189 is a .tiff image of a firebrand - notice how the middle of the firebrand shows up (red) in Windows Photo Viewer and does not show up (black) when opened in Matlab using imshow(). Any suggestions are appreciated!
Attach the TIFF image. It looks like they're using different colormaps. It's not an RGB image is it? Because it looks like probably it's an indexed image and the colormaps are different.
Hmm it appears to be too large: even when compressed into a zip file
How large are they? Can you subsample them to make them smaller?
This is the best I can do. Taken a super small crop of the initial .tiff and zipped it with Adobe Photoshop (which changed to .tif) and then into a zip folder. Original .tiff image is 118 MB.
I know it's too late, but I saw this while cleaning up. There's a bunch of Photoshop metadata in that file, the purpose of which I don't understand. There is an included ICC profile, but there's also a 1x4236 numeric vector embedded by Photoshop. So... more profile data?
Maybe in another five years, someone will tell us what it all means.

Sign in to comment.

More Answers (2)

A mitigation to a pixelated image display may be interpolation available from R2019b:
figure;
subplot(121); imshow('logo.tif')
subplot(122); imshow('logo.tif', 'Interpolation','bilinear')
I use it to display text pages converted to images.
My thanking to Steve Eddins:
meola22
meola22 on 2 May 2017
Edited: meola22 on 3 May 2017
Just stumbled upon the same problem and the solution is quite simple. All one needs to do is change the OpenGL from software to hardware. So just add this line of code:
opengl hardware
That fixed the pixelation problem for me.

1 Comment

This improved it, thx! It's still not 100% perfect resizing (as I get i photoshop), so I'm still working on it.

Sign in to comment.

Tags

Asked:

Ian
on 1 Dec 2014

Commented:

DGM
on 5 Feb 2023

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!